Federal Funding to Universities in the U.S. 2023

Author

Gaston Sanchez

Published

May 7, 2025

1 Description

The data visualization of this module is based on the following New York Times article by Andrea Fuller, Zach Levitt and Isabelle Taft (April 30, 2025):

Where Federal Dollars Flow to Universities

https://www.nytimes.com/interactive/2025/04/30/us/university-funding-research.html

Federal funding for research and development in 2023


1.1 Details (datasheet)

  • Topic(s):
    • Economics
    • Education
    • U.S.
    • Map
  • Data:
    • Size: medium data
    • Raw data in Excel file (.xlsx)
    • Also, required curating (via ChatGPT) a CSV file with geospatial locations and types of schools
  • Graphic:
    • Type: map with superimposed points
    • Interactive: no, but we can use "plotly" and "ggiraph"

1.2 R Packages

library(tidyverse) # ecosystem of data science pkgs
library(readxl)    # import excel files
library(sf)        # "simple features" (geospatial objects)
library(rnaturalearth) # maps data base
library(usmap)     # U.S. map with Alaska and Hawaii
library(leaflet)   # web interactive maps
library(ggiraph)   # make ggplot into interactive graphics

2 Data

The graphics in this module depend on 2 data files:

  • Excel: nsf25314-tab024.xlsx

  • CSV: highered-geo-type-nsf25314-tab024.csv

The main data set comes from the Higher Education Research and Development (HERD) Survey from 2023. This survey is the primary source of information on research and development expenditures at U.S. colleges and universities (link below).

https://ncses.nsf.gov/surveys/higher-education-research-development/2023#data

Specifically, the data comes from Table 24: Federally financed higher education R&D expenditures, ranked by FY 2023 R&D expenditures: FYs 2010–23. This data table can be downloaded as an Excel file: nsf25314-tab024.xlsx

In addition to the above Excel file, this module requires data about the location (longitude and latitude) and type (public or private) of the universities listed in the Institution Rankings table. We curated a CSV file by asking ChatGPT to provide the geospatial coordinates for the institutions in the rankings, and also the indicate whether they are public or private. This data is in the CSV file highered-geo-type-nsf25314-tab024.csv

2.1 Importing Data Files

To import the data in the Excel file nsf25314-tab024.xlsx use the read_excel() function as follows:

dat24_raw = read_excel(
  path = "data/nsf25314-tab024.xlsx", 
  sheet = 1, 
  col_names = FALSE)

# select columns: Name (1), Rank (2), and Expenditures-2023 (29)
dat24 = dat24_raw |> 
  select(1, 2, 29)

colnames(dat24) = c("Name", "Rank", "Exp2023")

# expenditure amount in millions of dollars
dat24 = dat24 |> 
  mutate(Amount = Exp2023 / 1000)


Use read_csv() to import the CSV file containing the geospatial coordinates, and the type of university:

# Input file: highered-geo-type-nsf25314-tab024.csv
# File curated with ChatGPT to get geospatial coordinates of institutions
# Adjust long-lat limits to avoid institutions outside U.S. States
edu = read_csv(
  file = "data/highered-geo-type-nsf25314-tab024.csv",
  show_col_types = FALSE) |> 
  filter(Longitude < 0, 
         Latitude > 19) |> 
  mutate(Color = case_when(
    Type == "Public" ~ "#cc823799",
    Type == "Private" ~ "#8c677399"
  ))

3 Graphics

We provide various ways to plot the map.

3.1 Map 1

# Option 1) using map from "usmap" which includes Alaska and Hawaii
# Important: Convert spatial data (e.g. lon, lat) to "usmap" projection
edu2 = usmap_transform(
  data = edu, 
  input_names = c("Longitude", "Latitude"))

ggplot() +
  geom_sf(data = us_map(), #fill = "#f7f6f0"
          fill = "#f8f8f8") +
  geom_sf(data = edu2, 
          color = "white",
          shape = 21,
          fill = edu2$Color) +
  theme_void()

3.2 Map 2

Option 2) another less ideal option is to use a map from "rnaturalearth"; this map does not contain Alaska and Hawaii.

us_states = ne_states(
  country = "United States of America", 
  returnclass = "sf")

gg = ggplot() +
  geom_sf(data = us_states) +
  coord_sf(xlim = c(-130, -65), ylim = c(25, 50)) +
  geom_point_interactive(
    data = edu, 
    aes(x = Longitude, y = Latitude, tooltip = Institution), 
    color = "tomato", alpha = 0.5) +
  theme_void()

girafe(ggobj = gg)

3.3 Map 3

A third option is to get an interactive map using the package "leaflet"

edu |> 
  inner_join(dat24, by = c("Institution" = "Name")) |> 
  leaflet() |> 
  addTiles() |> 
  addCircles(lat = ~Latitude,
             lng = ~Longitude,
             label = ~paste0(Institution, "; $", Amount),
             color = ~Color)


A better look can be obtained by creating better labels, and also by customizing the tiles, and the visual appearance of the points.

edu_dat24 = edu |> 
  inner_join(dat24, by = c("Institution" = "Name"))

# better labels
labels <- sprintf(
  "%s<br/>$%s", 
  edu_dat24$Institution,
  format(1000 * edu_dat24$Amount, big.mark = ",")) |> 
  lapply(htmltools::HTML)

# map
edu_dat24 |> 
  leaflet() |> 
  addProviderTiles("CartoDB.Positron") |> 
  addCircles(lat = ~Latitude,
             lng = ~Longitude,
             label = ~labels,
             color = ~Color, 
             radius = ~100 * Amount,
             weight = 2, fillOpacity = 0.5)

3.4 Map 4

The closest version to the desired map can be obtained with the following commands:

# Merge with points-polygon data
edu3 = edu2 |> 
  inner_join(dat24, by = c("Institution" = "Name"))

# Map
ggplot() +
  geom_sf(data = us_map(), fill = "#f8f8f8", color = "#d8d8d8") +
  geom_sf(data = edu3, aes(size = Amount), 
          color = "#DDDDDD", 
          fill = edu3$Color, 
          shape = 21) +
  #scale_size_area(breaks = c(500, 1000, 2000, 3000)) +
  scale_size_continuous(breaks = c(500, 1000, 2000, 3000)) +
  theme_void() +
  labs(title = "Federal funding for research and development in 2023")